Research Groups Materials Modelling



Angela Quadfasel

Group Manager Materials Modelling I


+49 241 80 95948





Alexander Krämer

Group Manager Materials Modelling II


+49 241 80 95954



The Materials Modelling groups deal with microstructure evolution during forming processes and in corresponding process chains. The groups utilize macroscopic as well as multi-scale modelling techniques to predict grain size, texture and precipitation.

The group Materials Modelling I in addition is concerned with modelling phenomena occurring at different types of interfaces. Notably the interaction between tool and work piece, e.g. during skin-pass rolling, as well as the bond formation during roll bonding of metals. In addition, unavoidable phenomena like wear in metal forming are a major point of concern.

The group Materials Modelling II in addition is concerned with digitization in metal forming. The group uses intelligent methods based on machine learning to efficiently aggregate and evaluate data. For predictive modelling, the group combines physical, inverse and databased approaches to match requirements.


Process Chain Modeling of an Aluminium Wrought Alloy in AMAP P1

Process chain and consortium in the AMAP P1 project Copyright: AMAP Process chain and consortium in the AMAP P1 project

Within the research cluster "Advanced Metals and Processes", short AMAP, in Aachen, the process chain of an automotive outer skin alloy (AA6016) was investigated within the framework of Project 1 "Process modeling of rolled and annealed aluminum strip with special properties for the automotive industry". A special feature of the project is not only the approach to investigate a process chain in industrial production, in the laboratory and by means of numerical models, but also the consortium of the participating companies. Within Project 1, three major aluminium producers (Aleris, Hydro and Novelis), which would otherwise be in competition with each other, Mubea as the automotive supplier and SMS Group as the mechanical engineering company, are working together. In addition to the industrial partners, the Institute of Physical Metallurgy and Metal Physics, short IMM, and the Institute of Metal Forming, short IBF, of the RWTH Aachen University are also involved.

For further information, please contact Angela Quadfasel.


Investigation of Skin-Pass Rolling With a Focus on Surface

Sketch of the skin-pass process with mill finish and EDT surface Copyright: IBF Sketch of the skin-pass process with mill finish and EDT surface

An important characteristic of rolled aluminium strips for use in the automotive outer skin is the surface quality. The topography of the surface and in particular the number of roughness peaks as well as the volume of closed lubrication pockets influence the success of the subsequent process steps deep drawing and painting.
The work carried out so far has investigated the relationship between the process kinematics of skin-pass rolling and the transfer mechanisms. For this purpose, the kinematics of a process model of flat rolling was transferred to a mesomodel to describe the surface imprinting. With regard to the imprinting of the surface, a good correspondence between simulation and experiment could be shown.
In the medium term, the numerical model is intended to enable a knowledge-based design of the skin-pass process for aluminium alloys, taking into account global and local influences.

For further information, please contact Angela Quadfasel.

Simulation of Surface Indentation at Aluminium Skin-Pass Rolling
Simulation of surface indentation at aluminium skin-pass rolling

Pass Schedule Design via Machine Learning

Interaction of influencing parameters during pass schedule design for rolling Copyright: IBF Interaction of influencing parameters during pass schedule design for rolling

The design of pass schedules for rolling is based on expert and empirical knowledge. This is due to each pass influencing all subsequent passes. Machine learning algorithms could provide an approach to automatize the design of pass schedules. By training them they can derive knowledge from data without needing an explicit mathematical formulation. As a proof of concept a data base has been generated using a fast process model. A neural network has then been trained to design pass schedules based on the data provided, if the initial and final state of the workpiece are provided as input. The boundary conditions are given by the universal rolling mill available at the IBF. The designed pass schedule fulfills all boundary conditions while exactly meeting the final state of the workpiece. Therefor the automatized design of pass schedules using machine learning algorithms seems feasible.

For further information, please contact Alexander Krämer.


Simulation of the Process Chain for a Turbine Disc

Turbine disc process chain and position in the engine Copyright: Leistritz, SMS, IBF Turbine disc process chain and position in the engine

The production of turbine discs for aerospace applications is characterized by very strict safety requirements including tight windows for the microstructure. The evolution of the microstructure therefor needs to be accounted for during the design of the process chain. Accordingly an online-coupling between StrucSim, a program calculating the microstructure, and the commercial finite element, short FE, Software Simufact was developed. This means that StrucSim is called during the FE Simulation and influencing its results. Subsequently the process chain was reproduced in FE Simulations and calculated using the online-coupling. Thereby the microstructure evolution was calculated for the whole workpiece along the process chain. This technique can be used to optimize processes or process chains regarding productivity or reproducibility in the future.

For further information, please contact Alexander Krämer.


Fast Process Models for Rolling

Single pass during rolling including force, temperature and microstructure evolution Copyright: IBF Single pass during rolling including force, temperature and microstructure evolution

Fast process models enable the accurate simulation of heavy plate rolling on the industrial and laboratory scale. Based on the pass schedule and material parameters it predicts the most important properties, such as force, temperature and microstructure, within seconds. Thus it has a wide range of applications, particularly in the field of design and optimization. With Industry 4.0 in mind, it has been coupled to a data base of industrial trials resulting in the ability to determine material parameters from just the measured forces. It has furthermore been coupled with machine learning algorithms to automatically design pass schedules for the universal rolling mill at the IBF. Fast process models are also being used for teaching and seminars, supplemented by a specially created graphical user interface. It allows students and seminar participants to develop an intuitive approach to the design, calculation and optimization of pass schedule as well as a detailed understanding of the underlying mechanisms.

For further information, please contact Alexander Krämer.

Fast Rolling Models for Pass Schedule Design
Fast rolling models for pass schedule design

Finite-Element Based Process Design for Fabrication of Metal Composites by Roll Bonding

FE model for simulating bond strength evolution during Roll Bonding Copyright: IBF, Hydro FE model for simulating bond strength evolution during Roll Bonding

Roll Bonding enables the production of composites with customized combinations of properties. In roll bonding, the bonding partners are permanently joined together by plastic deformation. The bond formation is a complex process influenced by material properties and process parameters. At IBF an Abaqus subroutine has been developed for computing the formation and failure of the bonds. In a DFG transfer project, this subroutine will be further improved to develop efficient process routes for new material combinations. With this subroutine and the Abaqus process model, Roll Bonding can now be mapped. The bond strength is calculated depending on the surface enlargement. The established bond can also loosen again due to unfavorable load condition after roll gap. The influences of parameters such as temperature and height reduction on the bond strength and the bonding status can now be simulated.

For further information, please contact Zhao Liu.

  Simulation of roll bonding

High Manganese Steel Crashboxes

Experimental and simulated high manganese steel crashbox Copyright: IBF Experimental and simulated high manganese steel crashbox

High manganese steels, short HMnS, have a high energy absorption potential due to their extraordinary combination of strength and formability. This qualifies HMnS as potential materials for crash relevant components in the automotive industry. However, the available elongations up to 70% are not reached in the crash of thin walled structures. In order to use HMnS for crash-relevant lightweight structures, various measures have to be taken. These include an adapted alloy design and the adjustment of a tailored microstructure with increased yield strength. Thus, a defined deformation behavior with maximum energy absorption should be achieved. Accompanying the experimental investigation of the optimal material properties, the crash behavior is predicted by multi- scale simulation. Therefore, a physical-based hardening model with input data from ab initio calculations is coupled with the FEM simulation.

For further information, please contact Angela Quadfasel.


Microstructure Simulation With StrucSim

Simulation of flow stress with StrucSim Copyright: IBF Simulation of flow stress with StrucSim

StrucSim is a program developed at the institute of metal forming to predict the microstructure evolution as well as the flow stress for hot forming processes. The challenge in hot forming processes, especially in multi-stage hot forming processes (process chains), is the description of the interaction between hardening and softening of the material. To overcome this challenge, the microstructure of the material is described by state variables, which develop depending on the process parameters (temperature, time etc.). Thus, quantities as the mean grain size or the recrystallized fraction can be calculated, and the flow stress can be derived for each time point. StrucSim is successfully used in several industrial and scientific projects. The extension of the functionality of the program, as well as the coupling to FE programs, such as Simufact, Abaqus etc. are ongoing work.

For further information, please contact Rajeevan Rabindran.

Microstructure Calculation with StrucSim for Rolling and Forging
Microstructure calculation with StrucSim for rolling and forging

Precipitation in Aluminium Alloys in AMAP Project P19

Process chain and consortium in the AMAP P19 project Copyright: AMAP Process chain and consortium in the AMAP P19 project

Due to their high strength, good formability, corrosion resistance and relative low density, age hardable aluminium alloys of the system Al-Mg-Si (AA6xxx) exhibit a great potential for many light weight applications such as body-sheets for the automobile industry. For these alloys the high strength is mainly obtained by the precipitation microstructure. Within the industrial manufacturing, the characteristics and the evolution of the precipitation microstructure are determined by the time-temperature profile of the thermo-mechanic process chain. Usually the final strength is obtained by age hardening at elevated temperatures in the last process step. In order to optimize the manufacturing process regarding the in-service properties of the workpiece, a coupled tool to simulate the microstructure evolution at low/high temperatures as well as the resulting mechanical properties is developed at the IBF in the course of the AMAP project P19.

For further information, please contact Fabrice Wagner.


Cold Rolling Strategies for Producing Magnetic-Optimized Electrical Steel Sheet in Energy-Efficient Electrical Drives

Multi-scale model for simulating texture evolution during cold rolling Copyright: IBF, IMM Multi-scale model for simulating texture evolution during cold rolling

One way to increase the efficiency of electric drives is to optimize the magnetic properties of the electrical steel used in the magnetic core. In order to quantify the influence of process parameters on these final properties and to create a scientific-theoretical basis for the development of low-loss electrical steel, an interdisciplinary DFG research group, FOR 1897, is working on the integrated process chain modeling. The main task of the IBF is to investigate and simulate the cold rolling process. Experimentally, the IBF will test different rolling strategies on the cold and hot rolling mill. A multi-scale model that includes a macroscopic finite element model and a microscopic crystal plasticity finite element model is created to compute the texture evolution, which makes it possible to determine the influence of different rolling strategies and initial states on the local texture development during cold rolling. By linking the sub-models, it enables model-based process design of low-loss electrical sheets for highly efficient electric drives.

For further information, please contact Xuefei Wei.


FepiM-Algorithm, flow curve determination through explicit pointwise inverse modelling

Iterative adjustment of the flow curve in a single increment Copyright: IBF Iterative adjustment of the flow curve in a single increment

Determining flow curves directly from experimental outputs sometimes is non-trivial, especially when the deformation is inhomogeneous due to friction for example. The state of art inverse methods for flow curve determination fit the experimental force displacement curves with finite element (FE) simulation, thereby considering the geometry changes during deformation. However, these methods are cost and time inefficient, as well as requiring a predefined mathematical equation describing the flow curve. The FepiM algorithm is an inverse FE based method and determines the flow curve as tabular data. The flow stress at each increments in the simulation is determined by matching the simulated and experimental force at the current displacement. Therefore, the flow curve can be determined as tabular data, eliminating the necessity of mathematical equations to describe the flow curve. In a corresponding research project, different strategies to predict the flow stress are investigated, such as heuristic or iterative approaches. The final target is to enable fast flow curve extraction from inhomogeneous conditions.

For further information, please contact Aditya Vuppala.


Predicting the texture in fast models

Texture development during cold rolling Copyright: IBF Texture development during cold rolling

A more resource-efficient E-mobility requires increased efficiency of electrical drives. Optimizing the texture to reduce iron losses is one possibility to achieve this goal. There are models that can predict the texture development. However, these models employ calculation times of several hours. Therefore, the goal of the ERS Seed Fund Projects is to develop a fast model to predict the texture within seconds. An FFT-Solver from DAMASK provides a basis for the development of a fast model. The solver uses the deformation gradient from a forming process and applies it to a representative volume element to calculate the texture. A method called Model Order Reduction can reduce the computational time by interpolating between snapshots. Snapshots are results calculated beforehand serving as reference points. The universal rolling mill at the IBF, Abaqus and other simulations tools serve as comparisons to validate the results.

For further information, please contact Christian Idzik.